I need to understand thermal runaway

I was under the impression that LEDs always need a resistor inbetween them and their power supply, lest they go into thermal runaway and die.

And yet, in my cheap 9-LED 3-AAA light there is no resistance, and they don't die at all.

Not even if I ditch the AAAs and use a low-internal-resistance LiIon cell, which should be able to output enough current to give the LEDs a quick and painful death.

What gives?

Edit: never mind, I figured it out myself.
With a LiIon cell there aren't enough volts to have thermal runaway in the first place. At 4 volts every LED gets about 45 ma, which is a significant overdrive but won't kill them instantly.

thermal runaway can only occur in applications with much higher current consumption or conditions where heat is conserved instead of dissipation into the surrounding environment. driving 3V LEDs at 4.5V with 3AAA wont kill it immediately because theres not enough current from AAAs to sustain the drive, which results in generally low runtime and it increases deterioration over time. using 9V batteries can also drive LEDs [orange/red that I know of] without frying the die.

http://www.dansdata.com/caselight.htm

thermal runaway is basically excessive heat from a point of origin causing the surrounding to heat up. if there are conductive wires with current passing through it, heat will allow the conductor's internal resistance to diminish, causing more current to pass through, which causes even more heat...this cycle continues until something goes and break the circuit thats causing the heat.
I think theres a thread on this somewhere :thinking:...oh well, theres wiki

Another faithful dan's data reader, I presume
Yeah, I guess I should have thought about that article of his...

Fallingwater;

Thermal runaway is caused by the drop in voltage accross an LED as it heats up. That, in turn, causes more current to flow thru the LED. That destroys the LED.

One 3-Cree house light I made, dropped in voltage as I watched it heat up.

Larry Cobb

thermal runaway will happen to high power LED more often(that is why they need heatsink when driven at or above spec)

simply said, when the die of the LED gets hotter, it lets more current in, which result in the die getting even hotter and the evil cycle continue till the LED is destroyed.

In solid state devices, there are junctions called P-N junctions, that allow for the solid state amplification. I'll try to give you a graphical representation.

PPPPPPPPPPNNNNNNNNN
PPPPPPPPPPNNNNNNNNN
PPPPPPPPPPNNNNNNNNN

The Ps and Ns represent different positively and negatively valanced doping of the silicon substrate. The N region has a proclitivity to produce electrons, and the P region has a proclitivity to receive electrons. At their junction, there is a shared electrical field that places some strain on both regions. There is a pull from the electrons at the N region so as their electrons displace toward the P region. Now, this field is what encourages the N region to give electrons and the P region to receive them.

PPPPPPPPPPNNNNNNNNN
PPPPPPPPPPNNNNNNNNN
PPPPPPPPPPNNNNNNNNN

This field increases as the voltage differential at either sides of the transistor increases, this most importantly relates to gating (I won't get into this). Now, this is the part that most directly answers your question - an increase of thermal energy at this juncture (whether caused by the electrical current in the device or ambient environmental conditions, or both) exasperates this increase in the force and size of this region greatly. So it may look like this.


PPPPPPPPPPNNNNNNNNN
PPPPPPPPPPNNNNNNNNN
PPPPPPPPPPNNNNNNNNN

Long story short, this stronger field facilitates greater flow of electrons across the silicon substrate. Have you ever wondered why solid-state silicon conductors are called semiconductors? It is because silicon is not a very conductive element, however, when this PN electrical field is present, silicon can become very conductive. Varying the level of conductivity is how solid-state devices work. So, greater heat leads to greater field - this increase in field leads to more electrical current across the substrate - this increase in current generates more heat - which increases the size and strength of this field region - which generates more current - which generates more heat - and so on till the transistor may be destroyed. Basically, a threshold needs to be reached for such a device to go into thermal runaway, so keep your devices cool

Fallingwater said:

Another faithful dan's data reader, I presume
Yeah, I guess I should have thought about that article of his...

Click to expand...

not really
Had to do a search on it via google for the thread:laughing: